Originally Posted by FlyingArrow
Even if there is a cap on point differential, if advanced stats (beyond the game scores) are used as part of the formula, then those sorts of efficiency stats will be affected by garbage time. Better to be efficient and run up the score to a 50pt win whenever possible.
|
Precisely. But we have no idea how legitimately beneficial it is to do just that, because we have no idea how much weight is applied to offensive and defensive efficiency. Is it a little or is it a lot? Is it 1/5th of the entire rating system based on the 5-section infographic the NCAA distributed, or are pieces weighted differently? Perhaps with the machine learning and AI those offensive and defensive efficiencies are weighted more heavily in January and February to account for teams "gaining steam".
The criticism of the NET has never been whether its a poor ranking system. It may be exponentially better than the RPI. Nobody even cares about the hit-and-miss first rankings this week where things dont often make sense b/c the games are yet to be fully interconnected. The concern all comes down to zero transparency, zero ability to understand the very system you're being evaluated on, zero ability to schedule a non-conference season based on those understandings, zero confidence in those orchestrating the rating system in having their own fundamental understanding of what's taking place to the point where they can effectively articulate it to you, and zero accounting or auditing of the system by those member schools to ensure that the proper data is being inputted and the proper results are being outputted.
The lack of detail and transparency is so bad we can't even tell how much better the #7 team is compared to the #11 team. Is the #7 team equally worse than the #3 team as they are better to the #11 team? Or do teams get bunched up? Maybe the gap between #7 and #8 is five times greater than #8 and #9. Who's to say? Who's to say otherwise?
Without a rating value, rank is irrelevant when it comes to computers. If this were simply a vote poll like the AP Top-25 it wouldnt matter because its all subjective. But like the RPI, the NET is supposed to take most or all of the subjectivity away from the discussion -- thats the entire point of the system to begin with -- to remove bias and evaluate based on calling balls and strikes.
The scary part is I dont think most ADs, coaches, and admins understand what's even at stake here. Its just flying over their heads and they are ignorant to their own ignorance. And the NCAA is more than willing to let them remain that way if it can reduce the number of questions being asked.
When the NCAA can't answer basic questions on their own ranking system, the system is un-credible and cannot be trusted. Not when billions of dollars are in the mix, conference affiliations, coaching jobs worth millions, and national TV appearances are contracted. If someone handed you a gun and said "trust me its not loaded", you'd still check for yourself. Someone's pinky swear isn't enough.